Robert Tibshirani

Bio: Robert Tibshirani is an academic researcher from Stanford University. The author has contributed to research in topics: Lasso (statistics) & Elastic net regularization. The author has an hindex of 147, co-authored 593 publications receiving 326580 citations. Previous affiliations of Robert Tibshirani include University of Toronto & University of California.

An evaluation of tumor oxygenation and gene expression in patients with early stage non-small cell lung cancers

Abstract: Background: To directly assess tumor oxygenation in resectable non - small cell lung cancers (NSCLC) and to correlate tumor pO2 and the selected gene and protein expression to treatment outcomes. Methods: Twenty patients with resectable NSCLC were enrolled. Intraoperative measurements of normal lung and tumor pO2 were done with the Eppendorf polarographic electrode. All patients had plasma osteopontin measurements by ELISA. Carbonic anhydrase-IX (CA IX) staining of tumor sections was done in the majority of patients (n = 16), as was gene expression profiling (n = 12) using cDNA microarrays. Tumor pO2 was correlated with CA IX staining, osteopontin levels, and treatment outcomes. Results: The median tumor pO2 ranged from 0.7 to 46 mm Hg (median, 16.6) and was lower than normal lung pO2 in all but one patient. Because both variables were affected by the completeness of lung deflation during measurement, we used the ratio of tumor/normal lung (T/L) pO2 as a reflection of tumor oxygenation. The median T/L pO 2 was 0.13. T/L pO2 correlated significantly with plasma osteopontin levels (r = 0.53, P = 0.02) and CA IX expression (P = 0.006). Gene expression profiling showed that high CD44 expression was a predictor for relapse, which was confirmed by tissue staining of CD44 variant 6 protein. Other variables associated with the risk of relapse were T stage (P = 0.02), T/L pO2 (P = 0.04), and osteopontin levels (P = 0.001). Conclusions: Tumor hypoxia exists in resectable NSCLC and is associated with elevated expression of osteopontin and CA IX. Tumor hypoxia and elevated osteopontin levels and CD44 expression correlated with poor prognosis. A larger study is needed to confirm the prognostic significance of these factors. © 2006 American Association for Cancer Research.

Exploring the nature of covariate effects in the proportional hazards model.

Abstract: We discuss an exploratory technique for investigating the nature of covariate effects in Cox's proportional hazards model. This technique features an additive term sigma p1 fj(chi ij), in place of the usual linear term sigma p1 chi ij beta j, where chi i1, chi i2,...,chi ip are covariate values for the ith individual. The fj(.) are unspecified smooth functions that are estimated using scatterplot smoothers. These functions can be used for descriptive purposes or to suggest transformations of the covariates. The estimation technique is a variation of the local scoring algorithm for generalized additive models (Hastie and Tibshirani, 1986, Statistical Science 1, 297-318).

Imputing Missing Data for Gene Expression Arrays

Abstract: The singular value decomposition offers an interesting and stable method for imputation of missing values in gene expression arrays. The basic paradigm is • Learn a set of basis functions or eigen-genes from the complete data. • Impute the missing cells for a gene by regressing its non-missing entries on the eigen-genes, and use the regression function to predict the expression values at the missing locations. ∗Depts. of Statistics, and Health, Research & Policy, Sequoia Hall, Stanford Univ., CA 94305. hastie@stat.stanford.edu †Depts. of Health, Research & Policy, and Statistics, Stanford Univ, tibs@stat.stanford.edu ‡Life Sciences Division, Lawrence Orlando Berkeley National Labs & Dept. of Molecular. and Cell Biology, University of California. Berk.; eisen@genome.stanford.edu; §Department of Biochemistry, Stanford University;pbrown@cmgm.stanford.edu ¶Department of Genetics, Stanford University;botstein@genome.stanford.edu

Classification by Pairwise Coupling

Abstract: We discuss a strategy for polychotomous classification that involves estimating class probabilities for each pair of classes, and then coupling the estimates together. The coupling model is similar to the Bradley-Terry method for paired comparisons. We study the nature of the class probability estimates that arise, and examine the performance of the procedure in simulated datasets. The classifiers used include linear discriminants and nearest neighbors: application to support vector machines is also briefly described.

Supervised harvesting of expression trees

Abstract: We propose a new method for supervised learning from gene expression data. We call it 'tree harvesting'. This technique starts with a hierarchical clustering of genes, then models the outcome variable as a sum of the average expression profiles of chosen clusters and their products. It can be applied to many different kinds of outcome measures such as censored survival times, or a response falling in two or more classes (for example, cancer classes). The method can discover genes that have strong effects on their own, and genes that interact with other genes. We illustrate the method on data from a lymphoma study, and on a dataset containing samples from eight different cancers. It identified some potentially interesting gene clusters. In simulation studies we found that the procedure may require a large number of experimental samples to successfully discover interactions. Tree harvesting is a potentially useful tool for exploration of gene expression data and identification of interesting clusters of genes worthy of further investigation.

Scikit-learn: Machine Learning in Python

Abstract: Scikit-learn is a Python module integrating a wide range of state-of-the-art machine learning algorithms for medium-scale supervised and unsupervised problems. This package focuses on bringing machine learning to non-specialists using a general-purpose high-level language. Emphasis is put on ease of use, performance, documentation, and API consistency. It has minimal dependencies and is distributed under the simplified BSD license, encouraging its use in both academic and commercial settings. Source code, binaries, and documentation can be downloaded from http://scikit-learn.sourceforge.net.

Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2

Abstract: In comparative high-throughput sequencing assays, a fundamental task is the analysis of count data, such as read counts per gene in RNA-seq, for evidence of systematic changes across experimental conditions. Small replicate numbers, discreteness, large dynamic range and the presence of outliers require a suitable statistical approach. We present DESeq2, a method for differential analysis of count data, using shrinkage estimation for dispersions and fold changes to improve stability and interpretability of estimates. This enables a more quantitative analysis focused on the strength rather than the mere presence of differential expression. The DESeq2 package is available at http://www.bioconductor.org/packages/release/bioc/html/DESeq2.html .

Regression Shrinkage and Selection via the Lasso

Abstract: SUMMARY We propose a new method for estimation in linear models. The 'lasso' minimizes the residual sum of squares subject to the sum of the absolute value of the coefficients being less than a constant. Because of the nature of this constraint it tends to produce some coefficients that are exactly 0 and hence gives interpretable models. Our simulation studies suggest that the lasso enjoys some of the favourable properties of both subset selection and ridge regression. It produces interpretable models like subset selection and exhibits the stability of ridge regression. There is also an interesting relationship with recent work in adaptive function estimation by Donoho and Johnstone. The lasso idea is quite general and can be applied in a variety of statistical models: extensions to generalized regression models and tree-based models are briefly described.

Going deeper with convolutions

Abstract: We propose a deep convolutional neural network architecture codenamed Inception that achieves the new state of the art for classification and detection in the ImageNet Large-Scale Visual Recognition Challenge 2014 (ILSVRC14). The main hallmark of this architecture is the improved utilization of the computing resources inside the network. By a carefully crafted design, we increased the depth and width of the network while keeping the computational budget constant. To optimize quality, the architectural decisions were based on the Hebbian principle and the intuition of multi-scale processing. One particular incarnation used in our submission for ILSVRC14 is called GoogLeNet, a 22 layers deep network, the quality of which is assessed in the context of classification and detection.

Deep Learning

Abstract: Deep learning is a form of machine learning that enables computers to learn from experience and understand the world in terms of a hierarchy of concepts. Because the computer gathers knowledge from experience, there is no need for a human computer operator to formally specify all the knowledge that the computer needs. The hierarchy of concepts allows the computer to learn complicated concepts by building them out of simpler ones; a graph of these hierarchies would be many layers deep. This book introduces a broad range of topics in deep learning. The text offers mathematical and conceptual background, covering relevant concepts in linear algebra, probability theory and information theory, numerical computation, and machine learning. It describes deep learning techniques used by practitioners in industry, including deep feedforward networks, regularization, optimization algorithms, convolutional networks, sequence modeling, and practical methodology; and it surveys such applications as natural language processing, speech recognition, computer vision, online recommendation systems, bioinformatics, and videogames. Finally, the book offers research perspectives, covering such theoretical topics as linear factor models, autoencoders, representation learning, structured probabilistic models, Monte Carlo methods, the partition function, approximate inference, and deep generative models. Deep Learning can be used by undergraduate or graduate students planning careers in either industry or research, and by software engineers who want to begin using deep learning in their products or platforms. A website offers supplementary material for both readers and instructors.